Wenjing SHI

1) Integrated image analysis method with deep learning methods and applied them to video datasets.
2) Over 6 years’ experience in image features extraction, human detection, object tracking, human recognition, and human activity detection.
3) Over 4 years’ experience working with Python (OpenCV, Tensorflow, Keras, and PyTorch), over 5 years’ experience working with MATLAB, and over 3 years’ experience with C++ (OpenCV and OpenGL).

2023/06-present

Applied Scientist, Amazon Web Services, Bellevue, WA

Using AWS SageMaker Multi-label & Image Semantic Segmentation to label images. Developing deeplearning models (DeepLabV3+, Segment Anything Model, DivoV2, etc.) for scene segmentation. Evaluating machine learning model’s performance periodically with the confusion matrix, recall, precision, and accuracy.

2022/05-2022/08

Applied Scientist Intern, Amazon Web Services, Bellevue, WA

Developed SegNet +VGG16 deep learning model for scene segmentation. Used AWS SageMaker Image Semantic Segmentation to label images. Implemented algorithms through AWS Lambda, ECS, and CloudWatch. Deployed deep learning model in production with AWS SageMaker and CloudFormation.

2021/09-2021/12

Research Assistant, The Mind Research Network, Albuquerque, NM

Developed Dlib and Face Alignment codes to detect human faces through the Docker platform on GPU. Applied a developed VGG model to the GENKI-4K dataset and hundreds of long-term private videos to classify facial expressions.

2017/09-2023/05

Project Assistant, Advancing Out-of-school Learning in Mathematics and Engineering (AOLME), Albuquerque, NM

Collected video datasets for research use and built human detection, human recognition, and talking detecting systems. Generated ground truth for human activity recognition by using Matlab Video Labeler app. Documented algorithms of image processing and simulated robot color detection, maze escaping, and obstacle avoiding through Python and Google Colab. Generated and installed color detection, maze escaping, and obstacle avoiding algorithms for GoPiGo Raspberry Pi Robot.

2017/01-2023/05

PhD (with Distinction) in Computer Engineering, University of New Mexico, Albuquerque, NM. GPA -3.81/4.0

Dissertation: Long-term Human Participation Detection Using A Dynamic Scene Analysis Model

2015/01-2016/12

M.S. in Electrical Engineering, University of New Mexico, Albuquerque, NM. GPA - 3.66/4.0

Thesis: Human attention detection using AM-FM representations

2010/09-2014/06

B.S. in Communication Engineering, Tianjin University of Commerce, P. R. China, Tianjin. GPA - 3.7/4.0